A recent study analyzing 10,000 keywords has uncovered significant inconsistencies in the results generated by Google’s AI Mode. Conducted by SE Ranking, this research highlights the minimal overlap between the sources used by AI Mode and those identified in traditional organic search rankings. The findings illustrate that the performance of AI Mode, when compared to Google’s AI Overviews and the top organic search results, is markedly volatile. According to the study, “the average overlap of exact URLs between the three datasets was just 9.2%,” emphasizing the unpredictable nature of AI Mode.
The study reveals that AI Mode frequently pulls differing results, raising questions about its reliability. To assess consistency, researchers executed the same set of 10,000 keywords through AI Mode three times in a single day. The results exhibited considerable variation, with 21.2% of cases showing no overlapping URLs at all across the three response sets. Domain-level consistency was slightly improved at 14.7%, suggesting that AI Mode may draw different pages from the same domains.
One of the most striking findings is the minimal overlap between AI Mode and traditional organic search results. Only 14% of the URLs retrieved by AI Mode matched those in the top 10 organic search results for the same queries. When analyzing domain-level matches, this overlap increased to 21.9%. Notably, in 17.9% of queries, AI Mode returned no overlap with organic URLs, indicating that its selections might operate independently from Google’s established ranking algorithms.
On average, each response generated by AI Mode contains 12.6 citations. The majority of these citations are presented in block link format (90.8%), followed by in-text links (8.9%) and AIM SERP-style links (0.3%), which mimic traditional search engine results pages (SERPs). Despite the observed volatility, certain domains consistently appeared throughout the tests. The top-cited sites included:
Indeed (1.8%) Wikipedia (1.6%) Reddit (1.5%) YouTube (1.4%) NerdWallet (1.2%)Additionally, Google properties were cited most frequently, comprising 5.7% of all links, primarily from Google Maps business profiles.
When comparing AI Mode to AI Overviews, researchers discovered an average URL overlap of only 10.7%, with domain overlap reaching 16%. This disparity suggests that the two systems function under different operational logics, even though both are AI-driven.
The high volatility of AI Mode results introduces both challenges and opportunities for search marketers. Given that results can fluctuate even for identical queries, tracking visibility becomes increasingly complex. However, this variability also opens new avenues for exposure. Unlike traditional search results, where a select few top-ranking pages often dominate, AI Mode appears to refresh its citations more frequently. This means that publishers with relevant, high-quality content may have a better chance of appearing in AI Mode answers, even if they do not rank within the organic top 10.
To effectively navigate this evolving landscape, SEOs and content creators should consider the following strategies:
Prioritizing domain-wide authority and topical relevance Diversifying content across trusted platforms Optimizing local presence through tools like Google Maps Monitoring evolving inclusion patterns as AI Mode continues to developFor further insights and detailed findings, be sure to explore the full study from SE Ranking.
